Abstract

A review of published data relating to A-bomb survivors has led to the conclusion that since they were based on the mortality experiences of five year survivors estimates of radiation effects should have been controlled for two opposing forces-namely, selective survival of exceptionally fit individuals during the period of heavy acute mortality and residual disabilities. Both effects were dose-related and beyond question, and the disabilities probably included the effects of incomplete repair of bone marrow damage. Therefore, in addition to differences between high and low dose being largely obliterated, there was probably distortion of cancer effects. The two opposing forces are clearly the reason why the change from the high mortality rates of 1945-6 to the low rates of the 1950s was not accompanied by a change from a position to a negative association with dose, and imperviousness to the residual disabilities is probably the reason why sudden deaths of previously healthy individuals (exemplified by suicides) were an exception to this rule. Finally, impairment of bone marrow function probably accounts for the early epidemic of myeloid leukaemia; the apparent absence of other cancers at this time, and the relatively high dose-related death rates for blood diseases other than leukaemia.